A net clinical benefit analysis of warfarin and aspirin on stroke in patients with atrial fibrillation: a nested case–control study

Table 2 Adjusted rate ratios of cerebrovascular outcomes associated with the use of warfarin and aspirin

	Cases/Controls	Crude RR	Adjusted RR (95 % CI)*
Ischemic stroke	5519/55,022
No use of any therapy	1513/15,499	1.00	1.00 (reference)
Current use of warfarin monotherapy	896/13,238	0.67	0.65 (0.59, 0.71)
Below therapeutic range (INR: <2)^†	63/667	0.95	0.93 (0.71, 1.22)
Within therapeutic range (INR: 2–3)^†	132/1838	0.71	0.69 (0.57, 0.83)
Above therapeutic range (INR: >3)^†	31/361	0.86	0.82 (0.57, 1.20)
Unknown therapeutic range^†	670/10,372	0.64	0.62 (0.56, 0.69)
Current use of aspirin monotherapy	2002/18,399	1.11	1.05 (0.98, 1.13)
Intracranial hemorrhage	689/6858
No use of any therapy	114/1365	1.00	1.00 (reference)
Current use of warfarin monotherapy	242/2214	1.41	1.29 (1.00, 1.68)
Below therapeutic range (INR: <2)^†	13/126	1.32	1.16 (0.62, 2.16)
Within therapeutic range (INR: 2–3)^†	34/356	1.25	1.13 (0.74, 1.72)
Above therapeutic range (INR: >3)^†	13/47	3.63	3.26 (1.67, 6.38)
Unknown therapeutic range^†	182/1685	1.39	1.29 (0.98, 1.69)
Current use of aspirin monotherapy	172/2210	0.97	0.92 (0.70, 1.19)

Abbreviations: RR: Rate ratio; CI: Confidence interval; INR: International normalized ratio.
*Adjusted for excessive alcohol use, smoking status, obesity, peripheral artery disease, myocardial infarction, previous cancer, prior bleeds, thromboembolic disorders, ACE inhibitor use, angiotensin receptor blocker use, antidepressant use, antipsychotic use, NSAID use, and statin use. The stroke model was additionally adjusted for CHADS₂ score, while the intracranial hemorrhage model was additionally adjusted for the components of that score.
^†Mutually exclusive categories among current users of warfarin monotherapy.
Note: Current users of other antithrombotic therapies or combinations, as well as past users are not displayed in the Table, but were included in the regression model to allow the proper estimation of treatment effects. These represent 1108 cases and 7886 controls for the ischemic stroke model, and 161 cases and 1069 controls for the intracranial hemorrhage model.

ISSN: 1471-2261